Journal of Electrical Systems and Information Technology

Table 3 CNN and LSTM parameters and performance for \({W}_{y}\) forecasting

From: A novel technique for forecasting the optimal production of a wind generator installed at any geographical point located within a very large area

CNN parameters		LSTM parameters
INPUT_WIDTH: 18 h
LABEL_WIDTH: 12 h
Filters	50	Units	160
Units	300
Learning_rates	\({10}^{-3}\)	Learning_rates:	\({10}^{-3}\)
RMSE	0.54 m/s	RMSE	0.39 m/s
R²	85%	R²	90%

Back to article page